IR

IRHAI Framework

Context Phenotypes Imaging Scoring Interventions DSEA Engine

Official White Paper Interactive Module • March 2026

Chronic Pancreatitis:
From Fibrosis to Function & Imaging

Chronic pancreatitis (CP) lacks a universally adopted imaging-driven classification system. This framework transforms imaging into structured, decision-grade outputs, enabling reproducible and transparent clinical guidance.

Created by Dr. Sharad Maheshwari MD - imagingsimplified@gmail.com

⚠️ The Clinical Gap

Clinicians think in terms of pain patterns, ductal obstruction, and functional loss. Radiologists report calcifications and atrophy without actionable synthesis. Result: Delayed interventions and inappropriate ERCP.

🎯 The DCPF Solution

The Deterministic Chronic Pancreatitis Framework integrates pathophysiology, multimodality imaging, and decision triggers to align radiology directly with intervention pathways.

Pathophysiology

The Two Dominant Phenotypes

CP is a continuum of injury resulting in structural damage. Identifying the specific morphologic pattern is critical, as it directly identifies the patient's primary pain driver and guides subsequent management.

🔗

Ductal Obstructive

Pain Driver: Ductal Hypertension

• → Dilated main pancreatic duct (>5–7 mm)
• → Intraductal calculi present
• → Upstream gland changes

🍂

Parenchymal Fibrotic

Pain Driver: Neuropathic + Inflammatory

• → Gland atrophy
• → Extensive fibrosis
• → Minimal or no duct dilatation

Diagnostic Approach

Modality Hierarchy & Strategy

No single modality answers all clinical questions. The DCPF strategy assigns specific roles to CT, MRI, EUS, and ERCP to build a comprehensive, actionable picture of the disease.

Relative diagnostic/therapeutic utility across CP features.

☢️ CT

Structural

Primary tool for assessing structural destruction. Best for identifying calcifications and evaluating complications (pseudocysts, splenic vein thrombosis).

Gold Standard

🧲 MRI + MRCP

Ductal Anatomy

Provides the definitive ductal roadmap. Essential for identifying the "chain of lakes" appearance and planning interventions.

🔬 EUS

Early Disease

Utilizes Rosemont criteria to detect early disease and subtle fibrosis. Limitation: Highly operator dependent.

🎥 ERCP

Therapeutic

Not indicated as a primary diagnostic tool. Reserved strictly for therapeutic actions like stone extraction and stenting.

Standardization

Modality-Specific Scoring Systems

Unlike acute pancreatitis, CP scoring is fragmented. Because the disease affects ducts and parenchyma at different rates, grading systems are strictly tied to specific imaging modalities.

Cambridge Classification

MRCP / ERCP

The gold standard for grading ductal severity.

• Grade 0 (Normal): Normal MPD & side branches.
• Grade 1 (Equivocal): Normal MPD; 1-2 abnormal side branches.
• Grade 2 (Mild): Normal MPD; ≥3 abnormal side branches.
• Grade 3 (Moderate): Abnormal MPD (irregular/dilated) + ≥3 abnormal side branches.
• Grade 4 (Severe): Grade 3 features plus large cavity (>10mm), macroscopic stones, or strictures.

Rosemont Criteria

EUS

The standard for early parenchymal detection.

Major Features

• • Stones (A)
• • Honeycombing (A)
• • Lobularity (B)

Minor Features

• • Cysts
• • Dilated MPD
• • Strands/Margins

Outputs: 1. Consistent with CP | 2. Suggestive of CP | 3. Indeterminate | 4. Normal

Functional Grading

Secretin-MRCP

Grades exocrine reserve via fluid progression over 10 mins.

• Grade 0: Fluid reaches jejunum (Normal).
• Grade 1: Reaches 3rd portion of duodenum (Mild).
• Grade 2: Reaches 2nd portion of duodenum (Moderate).
• Grade 3: Confined to duodenal bulb (Severe impairment).

CT Descriptive Staging

CT

No eponymous scoring system exists. Graded descriptively: Early/Mild (often normal), Moderate (atrophy + mild dilation), and Severe (calcifications, severe atrophy, "chain of lakes").

M-ANNHEIM System

Clinical + Imaging

The most comprehensive severity index (0 to IV). It combines imaging morphology (Cambridge) with clinical pain requirements and functional exocrine/endocrine loss.

🧮 Interactive Scoring Calculators

Cambridge Classification Calculator

Main Pancreatic Duct (MPD) Normal (≤ 3mm, smooth) Abnormal (Irregular or Dilated)

Abnormal Side Branches 0 1 to 2 ≥ 3

Cavity >10mm, Macroscopic Stones, or Strictures present

Cambridge Grade Grade 0 Normal

Rosemont Criteria Calculator

Major Criteria Hyperechoic foci with shadowing (Stones) [Major A] Lobularity with honeycombing [Major A] Lobularity without honeycombing [Major B]

Minor Criteria

Cysts Dilated MPD Hyperechoic margins Hyperechoic strands Non-shadowing foci

EUS Classification Normal ≤ 2 Minor features

Functional Grading (Secretin-MRCP) Calculator

Maximal Fluid Progression (at 10 mins post-secretin) Fluid easily reaches Jejunum Fluid reaches 3rd portion of duodenum Fluid reaches 2nd portion of duodenum Fluid confined to duodenal bulb

Exocrine Reserve Grade 0 Normal Function

Clinical Action

Intervention Pathways

This is where the DCPF differentiates. Radiological findings must directly trigger actionable clinical pathways. Explore the required imaging prerequisites for each intervention.

Extracorporeal Shock Wave Lithotripsy

Non-invasive fragmentation of obstructive ductal stones to facilitate clearance.

📌 Indications

• Large ductal stones (>5 mm)
• Obstructive duct physiology

📸 Imaging MUST Report

• → Precise stone size
• → Location (head vs. body)
• → Degree of duct dilatation

Endoscopic Therapy (ERCP)

Minimally invasive approach to relieve obstruction via stenting or extraction.

📌 Used For

• Stone extraction
• Stricture stenting

📸 Imaging MUST Define

• → Presence of dominant stricture
• → Duct continuity
• ⚠️ Exclude abrupt cutoff (Rule out cancer)

Surgical Intervention

Definitive structural management for refractory disease or suspected malignancy.

📌 Indications

• Refractory pain
• Failed endotherapy
• Suspicion of malignancy

🔪 Procedure Types

• →
  Puestow
  Drainage for diffuse ductal dilatation.
• →
  Whipple
  For head-dominant disease.

Active System

DSEA Decision Tree Engine

To enable transparent AI alignment (IRHAI), reports must capture structured data via the SM Protocol. Input the measurable radiological variables below to trigger deterministic clinical pathways, avoiding black-box predictions.

INPUT

SM Protocol Variables Radiology Output

D Duct diameter Normal (≤ 5 mm) Dilated (> 5 mm)

S Stones No Yes

SZ Size ≤ 5 mm > 5 mm

STR Stricture No Yes

A Atrophy No Yes

🚨 M Mass / Abrupt Cutoff No focal mass or cutoff Present (Suspicious)

dsea_decision_engine.sh

Active

Processing imaging logic gates...

1. Phenotype Mapping

output > Evaluating...

2. Intervention Pathway

action > Evaluating...

3. Malignancy Risk

risk > Evaluating...

IRHAI Logic Trace Log No Black-Box Predictions

irhai@system:~$

IR

Institute for Responsible Healthcare AI (IRHAI)

Based on the white paper "From Fibrosis to Function: Deterministic Imaging and Decision Framework for Chronic Pancreatitis" by Dr. Sharad Maheshwari MD - imagingsimplified@gmail.com.

Module Version 1.0 • Transparent AI Alignment